In a move expected to dramatically increase the quantity of available audiobooks, Spotify announced on Thursday that will begin accepting titles narrated by vocal AIs from ElevenLabs.
“For authors looking for a cost-effective way to create high-quality audiobooks, digital voice narration by ElevenLabs is a great option,” the company wrote in its announcement post, pointing out that AI narration has been one of its most requested features. “Authors can use the ElevenLabs platform to narrate their audiobooks in 29 languages, with complete control over voice and intonation.”
In order to submit a digitally-narrated book to Spotify, authors will first have to upload their work to ElevenLabs where the platform’s AI will generate a voiceover. Please note that unless you’re having it narrate a pamphlet, you’ll likely need to shell out for one of ElevenLabs subscriptions, which can range from $22 per month for the Creator tier to $1,320 per month for the Business tier. According to rival vocal AI startup, Speechify, an average-length audiobook runs for around 10 hours.
Once the AI has completed its generative task, authors can simply download the resulting file package before uploading that to Spotify’s audiobook distribution service, Findaway Voices. The submission will then be reviewed (presumably by humans and not just another AI) for accuracy before being published. Spotify will clearly label its AI-narrated offerings as such so that readers can easily identify them.
Whether established authors, many of whom are already rightfully wary of allowing AI systems to use their works as training fodder, will embrace the new service remains to be seen. Hugo and Locus award-winning author, John Scalzi, for example includes a rider in his contracts guaranteeing that only human narrators work on his books.
I have it as a contractual point that my narration has to be done by humans.
